The main objectives of this research program are to determine what functional sets of neurons make up the sensorimotor cortex of the domestic cat, how the different sets are interconnected to form the "neuronal circuitry" of the cerebral cortex, what thalamic inputs end upon what sets, which sets send information to what subcortical sites, and how activity in these cerebral "neuronal circuits" generates the patterns of external, net vertical current flow and surface voltage changes ("spontaneous" and evoked potentials) characteristic of the tissue. To attain these objectives, studies are made of the response properties of large numbers of individual neurons, observed through extracellular microelectrode recording, during maneuvers specially designed to characterize them functionally, and during each of eight different "spontaneous" and evoked potentials. Through the methods of population analysis, the "circuitry" of the tissue is determined and the characteristics of its operation during each of the "spontaneous" and evoked potentials is analyzed. For each potential, the pattern of net vertical current flow is calculated from the rate of change of the voltage field through depth in the tissue. The relationship of this pattern of current flow to the pattern of neuronal activity that produced it is then studied quantitatively, and the relationship of each to the surface-recorded potential is determined. From these studies, it is possible to determine whether sensorimotor cortex has only two preferred modes of operation, as now seems probable. To aid in attaining the above objectives, a fully automated research facility is under development. Using a computer-controlled microdrive system, neurons responsive to the hunting stimulus are automatically detected, isolated and identified, and a pre-defined dynamically altering, experiment is then performed on each neuron. Appropriate stimuli are delivered under computer command, data are acquired by the computer and immediately analyzed. Full development of the technique should increase the data yield per animal five to tenfold, with a uniform rather than varying sampling bias.